## 1. Houses Price Episodes

## Source details

**Canonical URL:** [1. Houses Price Episodes](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15246.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15246.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15246.pdf.json)

---

### Scope and topics covered
- Presents analysis focused on house price episodes and a related section on "Financial and Business Cycle Episodes".
- Emphasizes empirical episode-based analysis of house prices and links to public sector debt dynamics.
- Analytical components include episode identification and median dynamics analysis, comparison with financial and business cycle episodes, examination of public sector debt responses and biases, VAR estimates of shocks in real house price growth and induced public sector debt bias, decomposition of public sector debt bias across components, and cross-country comparisons of median debt bias.

### Key findings and summary results
- Real estate cycles produce a public sector debt bias of around 5–6 percent of GDP on average over the cycle for a symmetric house price shock (10 percent increase in upturns and 10 percent decrease in downturns).
- House price cycles produce a public sector debt bias roughly three times larger than bias from output cycles and other financial variables.
- Asymmetric impact of house price episodes on public debt:
  - Reduction in public debt during financial upturns is much smaller (about a third) than the increase during downturns—2 percent versus 6 percent of GDP (textual comparison).
- Episode counts identified (sample of 30 countries; 1975Q1 to 2013Q3):
  - House price episodes: 59 positive, 66 negative.
  - Private credit episodes: 44 positive, 45 negative.
  - Equity episodes: 55 positive, 71 negative.
  - Output gap episodes: 55 positive, 71 negative.
- Persistence and recovery:
  - House price growth episodes are more persistent than credit, equity or output growth episodes.
  - House price growth remains around 4.5 and 5.5 percentage points below pre-trough and pre-peak levels respectively after 10 quarters.
- Net increases in public debt from simple correlations across episodes:
  - House prices: 6 percent of GDP.
  - Private sector debt: 2 percent of GDP.
  - Stock market cycles: 1 percent of GDP.
- Distributional uncertainty in long-term VAR debt-bias estimates:
  - Long-term debt bias median: 5 percent of GDP.
  - Long-term debt bias 5th percentile: ≈ 3 percent of GDP.
  - Long-term debt bias 95th percentile: ≈ 13 percent of GDP.
  - Only 2.5 percent chance estimate is below 3 percent of GDP.

### VAR estimates and quantitative results
- Methodology highlights:
  - State-dependent panel VARs with baseline vector Z including house price growth (or output gap), real GDP growth, inflation rate, change in policy rates, and public debt in percent of GDP.
  - Cholesky IRFs normalized to a house price shock of 10 percent for comparability.
  - Stochastic simulations: 5000 simulated subsamples (random sample of 25 countries and 140 observations) used to extract median debt bias and 95th percentile confidence intervals.
  - One standard deviation in real house price growth = 3.9 percent.
- Baseline VAR (without dummies) responses to a one standard deviation real house price growth shock:
  - Real GDP growth: positive and significant impact, around 2 percent in the long-run.
  - Public debt: negative and significant impact, 4.7 p.p. (percent of GDP).
  - Inflation: positive and significant impact, around 2 percent.
  - Interest rates (policy rate): positive impact, 0.15 p.p.
- State-dependent results:
  - Negative house price shocks: large and positive impact on public debt.
  - Positive house price shocks: small, negative but barely significant impact on public debt.
  - Asymmetry remains across all states of the output gap.
- Quantitative VAR debt-bias estimates for a 10 percent price fluctuation:
  - 4 percent of GDP over 10 quarters (aggregate short-run).
  - 5 percent of GDP in the long-run across abnormal house price episodes (statistically significant).
- Debt bias for output gap episodes:
  - 0.7 percent of GDP over 10 quarters for a 10 percent shock (small and barely significant; significance mainly medium-term).

### Factors driving public debt bias
- Decomposition (median) and mechanisms:
  - Reduction in debt during house price upturns is mostly due to GDP growth; fiscal deficits continue to contribute to higher debt.
  - Overall balance responds more to output fluctuations, producing lower bias when output cycles are the driver.
  - Fiscal policy tends to be pro-cyclical during financial cycles (except stock market driven), and moderately countercyclical during output downturns.
  - Debt dynamics worsen during downturns due to vanished growth effects and increased balance sheet transfers (e.g., bailouts) or valuation effects.
- Role of initial debt levels and episode severity:
  - Debt bias is stronger when public debt is high or episodes are large/long.
  - When public debt is above 100 percent of GDP on average, it increases by close to 17 percent of GDP over house price cycle episodes (median summary).
  - VAR estimates for high public debt countries:
    - House price induced debt bias ≈ 10 percent after 10 quarters and 19 percent after 20 quarters.
  - Extreme and sustained episodes lead to stronger bias, mostly from larger negative growth impacts.
- Policy buffers and private/financial sector indebtedness:
  - Debt bias is small for countries with policy buffers.
  - High financial or non-financial private sector debt yields highest house price induced debt bias (VAR estimate ≈ 9 percent of GDP after 10 quarters).
  - Low private or public debt results in marginal debt bias; such countries save a larger fraction of windfalls during booms.

### Policy implications and recommendations
- Central conclusions for policy:
  - Debt bias over the cycle is significantly larger for house price cycles than for stand-alone business cycles.
  - Revenue windfalls during asset-price upswings are often treated as permanent, leading to permanent tax cuts/increased spending that are hard to reverse.
  - Countries with limited fiscal space (high private, financial, or public sector debt) are forced into pro-cyclical adjustments during downturns.
- Specific policy recommendations:
  - Refine structural fiscal balance measures to account for movements in financial cycles.
    - Ex ante structural fiscal targets should take debt bias into account given uncertainty in channels from financial cycles to fiscal stance.
  - Strengthen fiscal institutions and fiscal rules to smooth pro-cyclicality by preventing loose fiscal stances in good times and supporting credibility in bad times.
  - Allow automatic stabilizers to fully operate throughout the cycle in the absence of debt sustainability concerns; requires correct measure of structural balance accounting for financial and business cycles.
  - Internalize financial sector risks into fiscal buffers, weighted by likelihood and uncertainty of cycle identification; bailouts have disproportionate long-lasting fiscal effects.
  - Avoid policy instruments and tax policies that exacerbate debt bias associated with housing/financial cycles (De Mooij, 2011), including:
    - Tax advantages for corporate debt financing.
    - Tax preferences for owner-occupied housing.
    - Regulatory distortions (rent control, building constraints).
  - Prioritize macroprudential policies as first line of defense against financial booms and busts.

### Data, sample, and technical annex (high-level)
- Sample and data coverage:
  - Quarterly data for 30 advanced and emerging economies; geographically mostly Europe plus US, Canada, Australia, Japan.
  - Sample periods vary by country; examples include US sample: 1975Q1 to 2013Q2 and overall up to 2013Q3.
  - Data sources: Bank for International Settlements; Bloomberg; Eurostat; Haver; National Authorities; OECD House Price Index; IMF World Economic Outlook; IMF staff calculations.
- Technical annex summary:
  - Panel VAR variables: change in public debt-to-GDP ratio, real house price growth, policy rate change, real GDP growth, inflation rate change.
  - Nine states from combinations of house price (normal/high/low) and output gap (normal/high/low); historical probabilities provided in Table A.1 (see source).
  - Cholesky IRFs normalized for a 10 percent house price shock.
  - Lag selection: six lags chosen via BIC.
  - Stationarity: panel unit root tests indicate stationarity for all endogenous variables.
  - Pairwise Granger causality tests and diagnostics reported in appendix.
  - Stochastic simulations: 5000 simulated subsamples used to generate distributional measures (median, 95 percent confidence intervals, percentiles) for debt bias.

*Source: _wp15246 - 1. Houses Price Episodes (PDF chapter/section).*

### 1. Houses Price Episodes ...............................................................................................

### 1. Houses Price Episodes

### Scope and topics covered
- Presents analysis focused on house price episodes.
- Contains a related section on "Financial and Business Cycle Episodes".
- Emphasizes empirical episode-based analysis of house prices and links to public sector debt dynamics.

### Analytical components
- Episode identification and median dynamics analysis.
- Comparison of house price episodes with financial and business cycle episodes.
- Examination of public sector debt responses and biases associated with house price episodes.
- Use of VAR (Vector Autoregression) estimates to assess shocks in real house price growth and induced public sector debt bias.
- Decomposition of public sector debt bias across components.
- Cross-country comparisons of median debt bias for different groups of countries.

### Figures and empirical materials (listed in source)
- Median Financial and Business Cycle Episode Dynamics
- Median Real GDP around Financial and Business Cycle Episode
- Median Public Sector Debt Bias During Episodes
- Response to a One Standard Deviation Shock in Real House Price Growth
- Median House Price Induced Public Sector Debt Bias from VAR Estimates
- Public Sector Debt Bias
- Median Overall Balance During House Price Episodes
- Public Sector Debt Bias Decomposition, Median
- Median Debt Bias for Different Groups of Countries
- Debt Bias for Different Country Groupings

*Source: _wp15246 - 1. Houses Price Episodes (PDF chapter/section) — canonical source URL provided in the content unit.*

### References .............................................................................................................

### _wp15246 - References .............................................................................................................

### I. INTRODUCTION
- Public debt in advanced economies fell from an average of 140 to 30 percent of GDP after World War II until the mid 70s; it began rising again from the mid 70s and was significantly exacerbated by recent financial and euro area crises (Abbas et al, 2011).
- Financial cycles strengthen the interplay between credit, asset prices, and public finances, amplifying vulnerabilities via leverage (Minsky, 1964) and debt-deflation dynamics (Fisher, 1933).
- Three main channels through which financial cycles affect fiscal policy:
  - Direct: tax revenues (asset-price driven revenue swings).
  - Indirect/output: automatic stabilizers responding to output changes.
  - Balance sheet transfers: public support for banks (bailouts) (Eschenbach and Schuknecht, 2004).
- During real estate and construction booms fiscal balances improve (direct revenue and output channels); busts lead to deterioration through falling asset prices, slower growth, and bailouts.
- Financial cycles produce a public debt bias: increases in debt during downturns exceed reductions during upturns.
- Key empirical strategy (overview):
  - Identify financial upturns/downturns using house prices, private non-financial debt, and stock market prices; contrast with output upturns/downturns.
  - Estimate panel VAR allowing dynamic interaction between house prices, real GDP growth, and public debt while controlling for supply and demand shocks.
- Main findings (summary):
  - Real estate cycles result in a public sector debt bias of around 5–6 percent of GDP on average over the cycle for a symmetric house price shock (10 percent increase in upturns and 10 percent decrease in downturns).
  - The bias is stronger when debt—public, private or financial sector—is already high; it is smaller for countries with low private sector debt.
  - The risk of significantly higher debt bias than point estimates is much larger for financial cycles than for stand-alone business cycles, especially for countries with high private or public debt.
  - Automatic stabilizers and discretionary fiscal policy respond more symmetrically across the cycle when they do not correlate to financial cycles; business cycles not coinciding with real estate cycles result in a much smaller bias.
- Paper organization: Section 2 identification/stylized facts; Section 3 VAR estimates; Section 4 factors driving debt bias; Section 5 policy implications.

### II. STYLIZED FACTS OF FINANCIAL AND OUTPUT GAP EPISODES
- Episode identification rule:
  - A country-specific downturn (upturn) is identified if the decline (increase) in a variable is more than one standard deviation below (above) the country-specific mean for at least three consecutive quarters; dynamics analyzed 10 quarters before and after peak/trough.
- Sample:
  - 30 countries; 1975Q1 to 2013Q3 (data availability varies by country and variable).
- Episode counts identified (Table 1 summary):
  - House price episodes: 59 positive, 66 negative.
  - Private credit episodes: 44 positive, 45 negative.
  - Equity episodes: 55 positive, 71 negative.
  - Output gap episodes: 55 positive, 71 negative.
- Persistence and recovery patterns:
  - House price growth episodes are more persistent than credit, equity or output growth episodes.
  - Growth of private sector debt, stock market index and output gap fully recovers to pre-peak/pre-trough levels within ten quarters; house price growth does not.
  - House price growth remains around 4.5 and 5.5 percentage points below pre-trough and pre-peak levels respectively after 10 quarters (Figure 1).
- Volatility ranking (peak-to-trough): stock market episodes most volatile, followed by house price episodes, with output gap episodes least volatile.
- Synchronization of episodes:
  - Output recovers at least two quarters earlier than private sector debt but lags house prices during downturns.
  - Real GDP growth shows v-shaped pattern during negative output gap episodes.
  - During financial sector downturns (except stock-market driven), real GDP growth declines ~3.5 percentage points in the 10 quarters up to the trough and recovers to pre-trough levels within two years; post-trough growth remains positive during negative stock market episodes but at a lower level.
- GDP dynamics after booms:
  - After financial sector booms, real GDP growth declines gradually and persistently; GDP seems unchanged before and after episodes of high equity market growth but slows significantly post house price and private sector debt peaks.
- House price cycles and public debt bias:
  - House price cycles produce a public sector debt bias roughly three times larger than bias from output cycles and other financial variables.
  - Of 66 positive and 59 negative house price episodes identified, availability of public debt data reduces usable sample to 33 positive and 35 negative episodes.
  - 15 out of 33 positive episodes reflect run-up to the latest financial crisis; two thirds of negative episodes reflect crisis impact.
  - Asymmetric impact: reduction in public debt during financial upturns is much smaller (about a third) than the increase during downturns—2 percent versus 6 percent of GDP (textual comparison).
  - Net increases in public debt from episodes (simple correlations): house prices 6 percent of GDP, private sector debt 2 percent of GDP, stock market cycles 1 percent of GDP.

### III. VAR ESTIMATES
- Methodology:
  - State-dependent panel VARs estimate asymmetric dynamic response of public debt to financial cycles, controlling for supply and monetary factors and allowing regime changes via dummies for abnormal house price growth and output gap.
  - Baseline vector Z includes: house price growth (or output gap), real GDP growth, inflation rate, change in policy rates, and public debt in percent of GDP.
  - Formal baseline model: Z_ct = ... (see source for full equation).
  - Use of Cholesky IRFs normalized to a house price shock of 10 percent for comparability.
  - Stochastic simulations: 5000 simulated subsamples (random sample of 25 countries and 140 observations) to extract median debt bias and 95th percentile confidence intervals.
- Baseline VAR (without dummies) responses to a one standard deviation real house price growth shock (one standard deviation = 3.9 percent):
  - Real GDP growth: positive and significant impact, around 2 percent in the long-run.
  - Public debt: negative and significant impact, 4.7 p.p. (percent of GDP).
  - Inflation: positive and significant impact, around 2 percent.
  - Interest rates (policy rate): positive impact, 0.15 p.p.
- State-dependent results:
  - Negative house price shocks: large and positive impact on public debt.
  - Positive house price shocks: small, negative but barely significant impact on public debt.
  - Asymmetry remains across all states of the output gap.
- Quantitative VAR debt-bias estimates:
  - For a 10 percent price fluctuation, VAR debt bias amounts to:
    - 4 percent of GDP over 10 quarters (aggregate short-run).
    - 5 percent of GDP in the long-run across abnormal house price episodes (statistically significant).
  - Debt bias for output gap episodes:
    - 0.7 percent of GDP over 10 quarters for a 10 percent shock (small and barely significant; significance mainly medium-term).
- Distributional uncertainty:
  - Long-term debt bias at the 5th percentile ≈ 3 percent of GDP versus median 5 percent.
  - 95th percentile debt bias ≈ 13 percent of GDP (almost three times the median).
  - Only 2.5 percent chance estimate is below 3 percent of GDP.
- Fiscal impact comparison:
  - Increase in public debt when house prices fell is larger than what can be explained by output cycle fluctuations.
  - Revenue gains during house price upswings are largely treated as permanent (tax cuts/increased spending), producing pro-cyclical constraints when bust occurs.

### IV. FACTORS DRIVING PUBLIC DEBT BIAS
- Decomposition of debt dynamics (median):
  - Reduction in debt during house price upturns is mostly due to GDP growth; fiscal deficits continue to contribute to higher debt.
  - On average, fiscal balances improve in upturns slightly more than they deteriorate in downturns, but continuing fiscal deficits even during upturns produce debt bias.
  - Overall balance responds more to output fluctuations, producing lower bias when output cycles are the driver.
- Cyclicality of fiscal policy:
  - Fiscal policy is pro-cyclical during financial cycles (except stock market driven), moderately countercyclical during output downturns.
  - Debt dynamics worsen during downturns due to vanished growth effects and increased balance sheet transfers (e.g., bailouts) or valuation effects.
  - Fiscal policy tends to reduce debt during house price and private sector debt downturns (pro-cyclical) and increases debt during upturns (except stock market driven upturns).
- Role of initial debt levels and episode severity:
  - Debt bias, especially from house price episodes, is stronger when public debt is high or episodes are large/long.
  - When public debt is above 100 percent of GDP on average, it increases by close to 17 percent of GDP over house price cycle episodes (median summary).
  - VAR estimates: for high public debt countries, house price induced debt bias ≈ 10 percent after 10 quarters and 19 percent after 20 quarters.
  - Countries with high public debt are unable to reduce it during house price upturns; output-cycle induced bias is marginally larger for high public debt countries but not as large as house price-induced bias.
  - Extreme and sustained episodes lead to stronger bias, mostly from larger negative growth impacts.
- Policy buffers and private/financial sector indebtedness:
  - Debt bias is small for countries with policy buffers.
  - High financial or non-financial private sector debt yields highest house price induced debt bias (VAR estimate ≈ 9 percent of GDP after 10 quarters).
  - Low private or public debt results in marginal debt bias; such countries save a larger fraction of windfalls during booms.

### V. POLICY IMPLICATIONS
- Central conclusions for policy:
  - Debt bias over the cycle is significantly larger for house price cycles than for stand-alone business cycles.
  - Revenue windfalls during asset-price upswings are often treated as permanent, leading to permanent tax cuts/increased spending that are hard to reverse.
  - Countries with limited fiscal space (high private, financial, or public sector debt) are forced into pro-cyclical adjustments during downturns.
- Policy challenges: diagnostics of financial cycles, quantifying fiscal impact, and designing appropriate policy responses.
- Specific policy recommendations:
  - Refine structural fiscal balance measures to account for movements in financial cycles.
    - Ex ante structural fiscal targets should take debt bias into account given uncertainty in channels from financial cycles to fiscal stance.
  - Strengthen fiscal institutions and fiscal rules to smooth pro-cyclicality by preventing loose fiscal stances in good times and supporting credibility in bad times.
  - Allow automatic stabilizers to fully operate throughout the cycle in the absence of debt sustainability concerns; requires correct measure of structural balance accounting for financial and business cycles.
  - Internalize financial sector risks into fiscal buffers, weighted by likelihood and uncertainty of cycle identification; bailouts have disproportionate long-lasting fiscal effects.
  - Avoid policy instruments and tax policies that exacerbate debt bias associated with housing/financial cycles (De Mooij, 2011).
    - Examples: tax advantages for corporate debt financing, tax preferences for owner-occupied housing, regulatory distortions (rent control, building constraints).
  - Prioritize macroprudential policies as first line of defense against financial booms and busts.

### REFERENCES (select citations as listed in source)
- Abbas, S. M. Ali; Belhocine, Nazim; ElGanainy, Asmaa; Horton, Mark (2011), “Historical Patterns and Dynamic of Public Debt – Evidence from a New Database”, IMF Economic Review 59.
- Ashenfelter, Orley and Alan Krueger (1994), “Estimating the Returns to Schooling Using a New Sample of Twins” American Economic Review 84 (5).
- Ashenfelter, Orley and Cecilia Rouse (1998), “Income, Schooling and Ability: Evidence from a New Sample of Identical Twins”, Quarterly Journal of Economics 113 (1).
- Benetrix, Agustin and Philip Lane (2011), “Financial Cycles and Fiscal Cycles”, prepared for EUI-IMF Conference “Fiscal Policy, Stabilization and Sustainability”, Florence, Jun 6-7, 2011.
- Borio, Claudio; Disyatat, Piti; Juselius, Mikael (2013), “Rethinking potential output: Embedding information about the financial cycle”, BIS Working Papers No. 404.
- De Mooij, Ruud A. (2011), “Tax Biases to Debt Finance: Assessing the Problem, Finding Solutions”, IMF Staff Discussion Note SDN/11/11.
- Eschenbach, Felix and Ludger Schuknecht (2004), “Budgetary risks from real estate and stock markets”, pp. 315 in Deficits and Asset Prices.
- Fisher, Irving (1933), “The Debt-Deflation Theory of the Great Depression”, Econometrica.
- Girouard, Nathalie and Robert W. Price (2004), “Asset Price Cycles, ‘One-Off’ Factors and Structural Budget Balances”, OECD Economics Department Working Papers No. 391.
- Minsky, Hyman (1964), “Financial Crisis, Financial System and the Performance of the Economy”, Private Capital Markets, ed. Commission on Money and Credit.
- Poghosyan, Tigran; Mattina, Todd; Xue Li, Estelle (2015), “Correcting ‘Beyond the Cycle:’ Accounting for Asset Prices in Structural Fiscal Balances”, IMF Working Paper WP/15/109.
- Sims, Christopher A., and Tao Zha (2006), “Were There Regime Switches in U.S. Monetary Policy?”, American Economic Review 96 (1).
- (Full list of references reproduced in source.)

### APPENDIX: DATA AND TECHNICAL ANNEX (high-level)
- Data coverage:
  - Quarterly data for 30 advanced and emerging economies; geographically mostly Europe plus US, Canada, Australia, Japan.
  - Sample periods vary by country (examples in source). US sample: 1975Q1 to 2013Q2 in one description; elsewhere sample span up to 2013Q3.
  - Data sources: Bank for International Settlements; Bloomberg; Eurostat; Haver; National Authorities; OECD House Price Index; IMF World Economic Outlook; IMF staff calculations.
  - Variables include indebtedness of households, non-financial corporate, financial corporate, government; fiscal variables (revenues, expenditures, interest payments); private credit; house price and stock market indices.
- Technical annex (state-dependent panel VAR):
  - Variables in panel VAR: change in public debt-to-GDP ratio, real house price growth, policy rate change, real GDP growth, inflation rate change.
  - Nine states of the world from combinations of house price (normal/high/low) and output gap (normal/high/low); historical probabilities provided (Table A.1).
  - Cholesky IRFs normalized for a 10 percent house price shock.
  - Lag selection: six lags chosen via BIC.
  - Stationarity: panel unit root tests indicate stationarity for all endogenous variables.
  - Pairwise Granger causality tests and diagnostics reported in appendix (see source for details).
  - Stochastic simulations: 5000 simulated subsamples used to generate distributional measures (median, 95 percent confidence intervals, percentiles) for debt bias.

*Source: IMF staff working paper content unit _wp15246 (References and chapter content as provided).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15246.pdf_